Machine learning for the improvement of patient flow

Thesis

This thesis presents how machine learning can be used to improve the allocation and use of resources in hospitals, in particular with respect to patient flow. A deep learning method is proposed that predicts where in a hospital emergency patients will be admitted after being triaged in the ED. Such a prediction will allow for the preparation of bed space in the hospital for timely care and admission of the patient as well as allocation of resource to the relevant departments, including during periods of increased demand arising from seasonal peaks in infections. The problem is posed as a multi-class classification into seven separate ward types. A novel deep learning training strategy is created that combines learning via curriculum and a multi-armed bandit to exploit this curriculum post-initial training. We also show that there are certain signifying tests which indicate what space of the hospital a patient will use. In showing that prediction of location of admission in hospital for emergency patients is possible using information from triage in ED, a new way of training neural networks using a teaching reinforcement learning agent is also introduced. The properties and strategies of the teacher are investigated before a federated learning method is developed to allow for learning from multiple hospitals simultaneously. It is hoped that this work will be of value to healthcare institutions by allowing for the planning of resource and bed space ahead of the need for it. This in turn should speed up the provision of care for the patient and allow flow of patients out of the ED, thereby improving patient flow and the quality of care for the remaining patients within the ED.

Authors: El-Bouri, R

• DOI: 10.5287/ora-eqrv9qjzb
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: machine learning; meta learning; healthcare; patient flow; reinforcement learning
• Subjects: Patient Flow; Meta learning; Deep learning (Machine learning); Reinforcement learning